JP5936154B2 - Acoustic transducer having gap control structure and method of manufacturing acoustic transducer - Google Patents

Description

The present invention relates to an acoustic transducer and a manufacturing method thereof.
[Cross-reference of related applications]
This application is based on provisional application No. 61 / 470,384 filed on Mar. 31, 2011 with the title of the invention “Acoustic Sensor with Gap-Controlling Geometry and Method of Manufacturing an Acoustic Sensor”. To do. The entire contents of which are incorporated herein by this reference.

  Current trends in acoustic transducer technology are aimed at smaller microphones. At present, electret microphones based on charged thin films are used in most applications. However, these microphones have the problem of deteriorating after exposure to high temperatures. Capacitive MEMS microphones are gaining popularity because they can withstand the high temperatures experienced during solder reflow, resulting in reduced manufacturing costs. Piezoelectric MEMS microphones have been studied for over 30 years and can potentially combine the advantages of electret microphones and MEMS capacitive microphones in a cost-effective manner. However, piezoelectric MEMS microphones have long suffered from high noise levels due in part to diaphragm tension caused by thin film residual stresses. For example, diaphragm microphones are constrained at every edge, which causes high diaphragm tension that results in reduced sensitivity. Conventional cantilever designs, such as rectangular cantilever microphones, are also subject to residual stresses that are substantially free from the surrounding substrate, with a small amount of residual stresses moving up or down the cantilever beam. Bend to any of the bottom and pull away from the substrate surface. This bending causes an increase in the gap around the cantilever, reducing acoustic resistance and undesirably reducing low frequency sensitivity.

  Thus, there is a need in the field of piezoelectric MEMS acoustic transducers to create new and useful acoustic transducers that have low frequency sensitivity regardless of residual stress.

FIG. 1 is a schematic diagram of a typical conventional transducer. FIG. 2 is a microscopic plan view of a cantilever according to a preferred embodiment of the present invention. 3A, 3B, 3C, 3D and 3E are schematic views of first, second, third, fourth and fifth preferred embodiments of the present invention, respectively. FIG. 4 is a schematic perspective view of a portion of a transducer according to a preferred embodiment of the present invention. FIG. 5 is a flowchart showing a method for manufacturing a transducer according to a preferred embodiment of the present invention. 6A-6H are schematic illustrations of an exemplary transducer manufactured in accordance with one variation of the preferred embodiment method.

  The following description of preferred embodiments of the invention is not intended to limit the invention to those preferred embodiments, but to enable one of ordinary skill in the art of MEMS acoustic transducers to make and use the invention. Is intended.

[Acoustic transducer]
As shown in FIGS. 2, 3 and 4, the preferred embodiment transducer 100 can include a substrate 160 and a plurality of cantilevered beams 120 arranged in a gap control array, The beam 120 includes a gap-controlling geometry 130. The preferred transducer 100 structure and arrangement allows the gap resistance to be controlled, which allows control of the frequency below the frequency at which the microphone response “drops” or decreases. Transducer 100 is preferably a piezoelectric transducer, but may alternatively be a capacitive transducer, an optical transducer (eg, a photoacoustic sensor), or any other suitable transducer that suffers from cantilever stress. Good. The preferred transducer 100 is preferably an acoustic transducer, more preferably an acoustic sensor (ie, a microphone), but may alternatively be a voltage or voltage driven and used as a speaker. Preferred transducers 100 are preferably incorporated into consumer electronics such as cell phones, but for medical applications (eg, hearing aids), photoacoustic detection, ultrasound applications, or other transducer 100 based applications as sensors or speakers. It may be used. The preferred cantilever arrangement of the transducer 100 serves to limit the gap size to the separation distance between adjacent cantilever beams 120. This is in contrast to conventional designs where, after separation, the residual stress in the cantilever can greatly refract the beam after separation, sometimes as much as the distance between the substrate and cantilever. In the transducer 100 of the preferred embodiment, the proximity between the multiple cantilever beams 120 on the substrate 160 makes the residual stress between the beams 120 comparable. Similar stress profiles between adjacent cantilevers 120 result in similar cantilever curvature, thereby limiting the gap size to the separation distance between adjacent cantilevers 120.

As shown in FIG. 4, the preferred transducer 100 can include a substrate 160. The substrate 160 of the preferred transducer 100 serves to support the transducer 100 during the manufacturing process and to support the cantilever 120 of the transducer 100 during operation. In one variation of the preferred transducer 100, the substrate 160 is at least partially composed of silicon or any suitable silicon-based compound or derivative (eg, Si wafer, SOI, poly-Si on SiO ₂ / Si, etc.). Alternatively, the substrate 160 can be constructed at least in part from any suitable combination of fiberglass, glass or materials. In another variation of the preferred transducer 100, the substrate 160 is substantially removed from the active area of the cantilever 120 during the manufacturing process so that the cantilever 120 has maximum propagation and response.

  As shown in FIGS. 2, 3, and 4, the preferred transducer 100 can include at least one cantilever 120. More preferably, the preferred transducer 100 can include a plurality of cantilevers 120 or an array of cantilevers arranged in a suitable arrangement, for example as shown in the embodiment of FIGS. 3A-3E. The preferred cantilever 120 of the transducer 100 serves to convert sound pressure into an electronic signal. The cantilever 120 preferably includes a gap control structure 130 that minimizes the resultant gap between each cantilever 120 in a greater or greater number of arrangements. The gap control structure 130 of the cantilever 120 preferably includes a distal end 132 and a proximal end 134. In the preferred cantilever 120, the distal end 132 is much smaller than the proximal end 134, and the cantilever 120 tapers from the proximal end 134 toward the distal end 132. The proximal end 134 is preferably substantially connected to the substrate 160, but with the exception of the proximal end, the entire cantilever 120 is preferably expanded or reduced to remove the applied stress. , Substantially separated from the surrounding substrate 160.

  As shown in FIGS. 3A-3E, the cantilever 120 preferably has a substantially pointed gap control structure 130 so that the proximal end 134 is in the direction of propagation of the cantilever 120. However, it is significantly wider than the tip 132 in the direction perpendicular to it. For example, in one variation of the preferred transducer 100, the cantilever 120 shown in FIGS. 3A, 3B and 3C can have a generally triangular shape. In another variation of the preferred transducer 100, the cantilever 120 may comprise a sector or wedge shape 130 having a substantially curved proximal end 134 as shown in FIG. 3D. In another variation of the preferred transducer 100, the cantilever 120 can have a square shape as shown in FIG. 3E. In other variations of the transducer 100 of the preferred embodiment, the cantilever 120 may be of any suitable shape with the distal end 132 narrower than the proximal end 134 in a direction perpendicular to the propagation direction of the cantilever 120. Can have various shapes. Suitable alternative shapes include any type of polygon or sector, and each cantilever 120 having an array can have substantially the same shape or substantially different shapes. The length of the cantilever 120 is preferably adjusted to match the desired resonant frequency of the microphone, but may alternatively be longer or shorter. The proximal end 134 is preferably twice as long as its length, but may alternatively have any width that allows the desired outer shape 102 of the transducer 100 to be achieved. .

  The cantilever 120 is preferably formed from alternating piezoelectric layers and electrode layers 142. Piezoelectric layer 144 can serve to convert the applied pressure into a voltage, and electrode layer 142 can serve to transmit the generated voltage to an amplifier such as a JFET, charge amplifier, or integrated circuit. it can. Piezoelectric layer 144 preferably includes aluminum nitride (AlN) due to its CMOS compatibility, but alternatively, lead zirconate titanate (PZT), zinc oxide (ZnO), polyvinylidene fluoride (PVDF), niobium. Lead acid magnesium-lead titanate (PMN-PT) or any other suitable piezoelectric material may be included. The electrode layer 142 preferably comprises molybdenum (Mo), titanium (Ti), aluminum (Al) or platinum (Pt), but may alternatively comprise any other suitable electrode material. The cantilever 120 preferably includes two piezoelectric layers 144 interposed between three electrode layers 142. However, the cantilever 120 can include three piezoelectric layers 144 interposed between the three electrode layers 142, or only a total of three layers (first electrode layer 142, first piezoelectric layer 144, and upper electrode layer). 142) or any number of layers in any suitable permutation of electrode layer 142 and piezoelectric layer 144. Preferably, the cantilever 120 incorporates at least one piezoelectric layer 144 and one electrode layer 142. In one configuration example, the electrode layer 142 preferably covers approximately 2/3 of the area of the generally triangular cantilever 120 to minimize the noise level, but depending on the shape of the cantilever 120 A wider or lesser area of the cantilever 120 can be alternately covered. Each electrode layer 142 preferably defines only one independent electrode for each electrode layer 142, but the electrode layer 142 is patterned to define a plurality of independent electrodes for each electrode layer 142. It may be a thing. The electrode layers 142 are preferably connected in series with each other by metal wiring, but may be connected in parallel or may be connected in both series and parallel.

  As shown in FIG. 3, the preferred transducer 100 is configured such that substantially the same cantilever 120 is disposed so that the tip 132 is collected in a common region near the center of the acoustic transducer 100. Preferably, each free edge 136 of each cantilever 120 is parallel to the free edge 136 of an adjacent cantilever 120, respectively. Each proximal end 134 of the cantilever 120 preferably forms an outer periphery 102 in the form of a regular polygon and / or a circle. The outer periphery 102 of the transducer 100 is preferably square, such that the transducer 100 preferably includes four cantilevers 120 (shown in FIG. 3A), but alternatively, the outer periphery 102 is an acoustic transducer (see FIG. A circle that includes any suitable number of wedge-shaped cantilevers 120 (shown in 3D), or a triangle in which the transducer 100 preferably includes three cantilevers 120 (shown in FIG. 3B), or An octagon such that the transducer 100 preferably includes eight cantilevers 120, or a hexagon such that the transducer 100 preferably includes six cantilevers 120 (shown in FIG. 3C), or a cantilever 120. It may be any geometric shape including any required number.

  In preferred transducers, the gap between the cantilevers 120 is less than about 1 micron during manufacture, but may be slightly larger. After manufacturing, the gap between the cantilevers 120 is preferably maintained below 1 micron, but can be much larger due to deformation caused by residual stress. The cantilevers 120 are preferably electrically connected via one or more electrode layers 142, but alternatively are electrically connected by conductive wiring 146 or electrically connected to each other. Or a combination thereof, that is, a portion of the cantilever 120 is electrically connected while the others are electrically isolated. The cantilevers 120 may be connected in series or connected in parallel, but are preferably connected in an extreme mixture of the two, ie, the cantilever 120. Some of them may be connected in series, while others may be connected in parallel.

[Acoustic transducer manufacturing method]
As shown in FIGS. 5 and 6, a preferred method for manufacturing a transducer includes a step of alternately depositing piezoelectric layers and electrode layers on a substrate in block S100, and a step of processing the deposited layers in block S200. The step of defining a cantilever shape, depositing metal wiring in block S300, and releasing the cantilever 120 from the substrate 100 in block S400 may be included. Since the transducer 100 is preferably assembled using a standard CMOS process, the associated electronics (eg, JFET, charge amplifier, integrated circuit) are assembled on the same substrate as the transducer 100 using the same CMOS process. be able to.

Block S100 of the preferred method describes depositing alternating layers of piezoelectric and electrodes on the substrate. Block S100 preferably functions to generate a cantilever layer. The piezoelectric layer preferably comprises aluminum nitride (AlN) due to its CMOS compatibility, but alternatively lead zirconate titanate (PZT), zinc oxide (ZnO), polyvinylidene fluoride (PVDF), niobic acid Lead magnesium-lead titanate (PMN-PT) or any other suitable piezoelectric material may be included. The electrode layer preferably comprises molybdenum (Mo), titanium (Ti), aluminum (Al) or platinum (Pt), but may alternatively comprise any other suitable electrode material. The cantilever is preferably manufactured using surface micromachining, but may alternatively be manufactured by bulk micromachining. Each layer is preferably deposited on the previous layer (the first layer is deposited on the SiO ₂ layer) and then etched to the desired pattern before depositing the next layer. Each layer is preferably deposited by thin film deposition, but may alternatively be deposited by reactive physical vapor deposition, physical vapor deposition, chemical vapor deposition, epitaxy, or any suitable method. Each layer is preferably first patterned by photolithography and then micromachined to remove material in areas exposed by photolithography. Micromachining methods include wet etching (chemical etching) and dry etching (eg, via reactive ion etching or ion milling), but may include any other suitable etching method. In one embodiment, every other electrode layer is formed by patterning the electrode layers so that the alternating layers are arranged in a staggered fashion (as shown in FIG. 6) and deposited in S300. It may be connected in parallel. However, the electrode layer and the piezoelectric layer may be patterned in any appropriate pattern.

  Block S200 of the preferred method describes processing the deposited layer to define a cantilever shape. Block S200 functions to generate a gap that defines the gap control structure of the cantilever. The deposited layer is preferably processed by etching a gap through the deposited layer (eg, by reactive ion etching, wet etching, ion milling or any other etching method), but alternatively In particular, the cantilever 120 may be defined and released from their neighbors to be processed in another way. The gap thickness is preferably 1 micron or less, but may alternatively be slightly larger. Also, the gaps preferably diverge from each other to form a substantially triangular cantilever, but alternatively may intersect at the ends to form the desired gap control structure. Good. This step preferably generates at least two bifurcated gaps to form at least four triangular cantilevers, but alternatively generates three, four or any number of gaps, Any number of cantilevers may be formed.

  Block S300 of the preferred method describes depositing metal wiring. Block S300 preferably functions to electrically connect the acoustic transducer to one or more amplifiers. Block S300 may be performed before, after, or simultaneously with block S200. The metal wiring is preferably patterned after being deposited as a single layer, but may alternatively be pre-patterned and deposited on the acoustic transducer. The block S300 preferably provides a metal wiring to each electrode or each electrode layer 142, but provides a single metal wiring to a plurality of electrodes so that the electrodes are connected in parallel to each other. Also good. The metal wiring preferably extends through the intermediate piezoelectric layer and / or electrode layer to the associated electrode layer 142, but may alternatively be connected to the transducer electrode in any suitable manner. Good.

  Block S400 of the preferred method describes releasing the cantilever from the substrate. Block S400 preferably functions to allow the cantilever to expand, contract, or bend as needed to substantially release residual stress. The cantilever is preferably released from the substrate by removing the substrate from below the cantilever. This is preferably achieved using DRIE (Deep Reactive Ion Etching), but wet etching, EDM (Electric Discharge Machining), micromachining process, or any other that releases the cantilever from the substrate. It may be achieved using a processing method. Alternatively, the cantilever may be reattached to either the same substrate or a different substrate after it has been completely released from the substrate. The cantilever is completely removed by providing a sacrificial layer between the substrate and the cantilever before beam layer deposition (ie, before block S100), and then etching away the sacrificial layer in block S500. Can be released. The sacrificial layer is preferably an oxide, but may be any suitable material different from the material of the piezoelectric and electrode layers that can be selectively removed. The sacrificial layer is preferably removed by an etchant such as hydrogen fluoride (HF) in an aqueous solution, by plasma etching, or by any other suitable etching process. The cantilevers are preferably reattached to the substrate along their proximal ends by electrostatic clamping or any suitable technique.

The preferred method may additionally include growing an oxide layer on the substrate at block S500. Block S500 is preferably performed before block S100 and preferably functions to control the amount of cantilever release in S400. In one variation of block S400, the substrate removal process preferably ends with an oxide layer. In another variation of block S500, the oxide layer preferably functions as a sacrificial layer. A suitable oxide is preferably grown on the desired active area of the transducer, but alternatively it may be grown in the desired release area of the transducer, over the entire substrate, or in any suitable area. It may be a thing. The oxide is preferably an oxide grown from a substrate, more preferably silicon dioxide (SiO ₂ ), but may be any suitable oxide grown or deposited on the substrate. The oxide is preferably grown using conventional thermal oxidation, but alternatively, plasma enhanced chemical vapor deposition (PECVD oxide deposition), chemical vapor deposition (CVD oxide deposition), physical vapor deposition It may be grown using (PVD oxide deposition) or any other suitable oxidation or oxide deposition process. The preferred method may additionally include removing the oxide layer in block S500A where the oxide layer is removed from the transducer by etching or micromachining.

  The preferred method may additionally include depositing a seed layer at block S600. The seed layer preferably functions as an active layer on which the cantilever is assembled. Block S600 is preferably performed before block S100. More preferably, block S600 is performed after block S500 such that a seed layer is disposed between the piezoelectric or electrode layer of the cantilever and the oxide layer. The seed layer is preferably aluminum nitride (AlN), but may be any suitable piezoelectric, electrode or seed material. The seed layer is preferably sputtered using physical vapor deposition (PVD) or any other suitable sputtering technique, but is otherwise deposited on the oxide layer or substrate. There may be.

[Method and Transducer of Example]
As shown in FIGS. 6A-6H, one implementation of the preferred method is to grow a thermal oxide (SiO ₂ ) on the substrate in block S500 and deposit an aluminum nitride (AlN) seed layer in step S600. Step (FIG. 6A), depositing and patterning a first electrode layer (molybdenum) (FIG. 6B), depositing and patterning a first piezoelectric layer (AlN) (FIG. 6C), and a second electrode layer Evaporating and patterning (mobidium) (FIG. 6C), evaporating and patterning the second piezoelectric layer (AlN) (FIG. 6D), and evaporating and patterning the upper electrode layer (molybdenum) in step S100. (FIG. 6D). Preferably, the cavity can be etched through the piezoelectric layer (AlN via) and into the electrode layer, and metal wiring can be deposited in block S300 (FIGS. 6E and 6F). In one variation, two metal lines are deposited, the first cavity / metal line extends to and is connected to the upper and lower electrodes, and the second cavity / metal line extends to the intermediate electrode to the intermediate electrode Connected. The cantilever is defined from the deposited layer S200 (etched or micromachined) (FIG. 6E) and from the substrate by etching the substrate from the back side with deep reactive ion etching (DRIE) in block S400. It is released (FIG. 6G). The DRIE stops at the oxide layer, which is removed to release the transducer from the substrate S500A (FIG. 6H).

  Throughout the execution of the example method, the residual stress is preferably monitored using wafer curvature measurements (eg, via optical or physical measurements), but alternatively the stress Measured by a measurement (eg stress transducer), non-linear elastic stress measurement (eg ultrasonic or magnetic techniques, X-ray or neutron diffraction) or any other method of measuring the residual stress or curvature of a cantilever beam There may be. The deposition parameters are then preferably adjusted to minimize cantilever deflection or stress.

  Those skilled in the art can now make modifications and changes to the preferred embodiments of the present invention from the foregoing detailed description and drawings, and from the claims without departing from the scope of the invention as defined in the following claims. You will recognize.

Claims (18)

A MEMS transducer comprising:
A substrate,
A plurality of tapered transducer beams,
Each tapered transducer beam includes a piezoelectric layer that converts an applied pressure into a voltage, and a pair of electrode layers that sandwich the piezoelectric layer therebetween,
Each tapered transducer beam has a beam base end, a beam end, and a beam main body disposed between the beam base end and the beam end, and the beam main body extends from the beam base end to the beam. Taper towards the edge,
The tapered transducer beam is connected to the substrate in a cantilever configuration by attaching the beam base end portion to the substrate, and the beam end portions of the plurality of tapered transducer beams are gathered toward a common point. The body and beam ends are not bonded to the substrate ,
2. The MEMS transducer according to claim 1, wherein the beam end portion includes a pointed beam tip, and the beam tips are gathered at almost one point . The MEMS transducer according to claim 1, wherein
A MEMS transducer, wherein the gap defined along the length of adjacent tapered transducer beams is about 1 micron or less. The MEMS transducer according to claim 1, wherein
A MEMS transducer, wherein each tapered transducer beam has a generally triangular surface. The MEMS transducer according to claim 1, wherein
A MEMS transducer, wherein at least two of said tapered transducer beams are electrically connected in series with each other. The MEMS transducer according to claim 1, wherein
The MEMS transducer, wherein the piezoelectric layer extends over substantially the entire width and length of the tapered transducer beam. The MEMS transducer according to claim 5 , wherein
Each of the tapered transducer beams includes two piezoelectric layers and three electrode layers. The MEMS transducer according to claim 6 , wherein
A MEMS transducer, wherein the electrode layers of each tapered transducer beam are electrically connected in series by metal wiring. The MEMS transducer according to claim 5 , wherein
The MEMS transducer, wherein the piezoelectric layer includes aluminum nitride, and the electrode layer includes molybdenum. Depositing a layer on which at least a first electrode layer, a piezoelectric layer and a second electrode layer are alternately overlapped on a substrate;
Processing the deposited layer to define a plurality of tapered transducer beams, each taper transducer beam converting the applied pressure into a voltage and the piezoelectric layer sandwiching the piezoelectric layer therebetween; Each taper transducer beam comprising a beam base end, a beam end, and a beam main body disposed between the beam base end and the beam end. When,
Depositing metal wiring;
Patterning the metal wiring;
Releasing the tapered transducer beam from the substrate, wherein the released tapered transducer beam is coupled onto the substrate in a cantilever configuration by attaching the beam proximal end to the substrate, Each beam end of the transducer beam gathers toward a common point, each beam body and beam end are not coupled to the substrate ,
The beam end portion includes a pointed beam tip, and the beam ends are gathered together at approximately one point . The method of claim 9 , wherein
A method wherein adjacent tapered transducer beams define a gap extending in the thickness direction of the deposited layer. The method of claim 10 , wherein
A method wherein the gap width between adjacent tapered transducer beams is about 1 micron or less. The method of claim 10 , wherein
A method wherein the gap is machined by micromachining. The method of claim 10 , wherein
The method wherein the tapered transducer beam is defined by two intersecting linear gaps. The method of claim 9 , wherein
The method of depositing the metal wiring includes depositing metal wiring on each electrode layer. The method of claim 9 , wherein
The method of releasing the beam from the substrate includes etching the substrate to separate it from the beam. The method of claim 15 , wherein
Etching the substrate comprises deep reactive ion etching. The method of claim 16 , wherein
A method further comprising growing a thermal oxide layer on the substrate prior to depositing and patterning the piezoelectric and electrode layers. The MEMS transducer according to claim 1, wherein
3. The MEMS transducer according to claim 1, wherein the beam ends are gathered toward one point without contact.

Images